A generic quantitative relationship between quality of experience and packet loss in video streaming services

Abstract

Increasing numbers of Internet users have recently been gaining interest in streaming services as they watch their favorite live football match or listen to the latest breaking news online through voice or video streaming services. It is known that in streaming services, even a small amount of quality disturbance can irritate users, in the form of, for example, blurred pictures or a lost scene. This sort of problem results in users becoming dissatisfied, and potentially giving up a service or even switching to another network or service provider. In the current competitive market, providers can thus incur large losses. To avoid such an issue, end user quality of experience (QoE) needs to be maintained at a satisfactory level, and to do so, knowledge of how Quality of service (QoS) parameters can affect QoE is vital. In this paper, a generic quantitative relationship between video streaming QoE and packet loss, as a QoS parameter, is proposed. The proposed relationship is then used to explain how user experience of video streaming quality changes in relation to fluctuations in QoS parameter. Data was collected from a test-bed in order to test the proposed relationship. Statistical techniques were also employed to show how successfully this relation can illustrate the dependency between video streaming QoE and packet loss. In addition, results show that functions in power forms, along with functions in exponential and logarithmic forms, can be considered candidates for assessing QoE based on QoS parameters.